Mean temperature calculation (day, season, most, comparison)

[ADGreen]

I am trying to calculate the mean temperature for any given location based on just the minimum and the maximum for the day (as I am unable to obtain the mean from my data source).

How good of an estimate can be obtained simply by taking the mean of the min and max?

For example, if the minimum is 8C and the maximum 24C, the estimated mean would be 16C.

I guess this depends on a number of factors (such as the location's humdity, season, elevation etc etc), but if there is a generic formula I could use as a rough guide, that would be good.

I really want the mean to reflect the mean temperature throughout the 24 hours of the day - it could be skewed more towards the minimum especially as the temperature may remain more stable at night (and cooler) and quickly heat up in the afternoon to reach the maximum.

So....is there a formula or really are there generic weighting factors that can be applied to both the min and max to get an approximate mean, ie

mean temperature = w1*min + w2*max

where w1 is the minimum weight and w2 is the maximum weight

[ChesterNZ]

The mean of the daily maxima and minima tends to be a little higher than the true mean (by about 0.2 - 0.6 C).

There is some interesting info on this subject in von Hann's "Handbook of Climatology" (partially available on Google books). He suggests calculating the mean as (min temp + max temp + 9 AM temp + 9 PM temp)/4.

[ADGreen]

Originally Posted by ChesterNZ

The mean of the daily maxima and minima tends to be a little higher than the true mean (by about 0.2 - 0.6 C).

There is some interesting info on this subject in von Hann's "Handbook of Climatology" (partially available on Google books). He suggests calculating the mean as (min temp + max temp + 9 AM temp + 9 PM temp)/4.

Thanks

I'll look out for that book as well - suspect I can get 9am and 9pm data which would appear to give a fairly close estimate (considering the simple mean is closer than I expected to true mean).

[RWood]

Originally Posted by ADGreen

Thanks

I'll look out for that book as well - suspect I can get 9am and 9pm data which would appear to give a fairly close estimate (considering the simple mean is closer than I expected to true mean).

If you're doing national or international intercomparisons you'll still have to stick to the simple (max+min)/2. For the purposes of year to year comparisons and a variety of other things, it's OK.

[Shaheen hassan]

Use the formula
(Max+Min)÷2

[Chaminade]

Originally Posted by ADGreen

I am trying to calculate the mean temperature for any given location based on just the minimum and the maximum for the day (as I am unable to obtain the mean from my data source).

How good of an estimate can be obtained simply by taking the mean of the min and max?

For example, if the minimum is 8C and the maximum 24C, the estimated mean would be 16C.

I guess this depends on a number of factors (such as the location's humdity, season, elevation etc etc), but if there is a generic formula I could use as a rough guide, that would be good.

I really want the mean to reflect the mean temperature throughout the 24 hours of the day - it could be skewed more towards the minimum especially as the temperature may remain more stable at night (and cooler) and quickly heat up in the afternoon to reach the maximum.

So....is there a formula or really are there generic weighting factors that can be applied to both the min and max to get an approximate mean, ie

mean temperature = w1*min + w2*max

where w1 is the minimum weight and w2 is the maximum weight

Many good weather sites provide the mean temperature for a location, most accurately if the data are from a local airport or other high quality station.